Authors: Agniva Das, Kunnummal Muralidharan

Published on: May 18, 2024

Impact Score: 7.8

Arxiv code: Arxiv:2405.11213

Summary

• What is new: The novel aspect of this research is the Hybrid Holt’s Model embedded with a Wavelet-based ANN, offering a new approach to forecasting COVID-19 dynamics compared to typical models.
• Why this is important: The need to predict the trajectory of COVID-19 cases, recoveries, and deaths accurately to better manage healthcare resources and governmental actions.
• What the research proposes: A Hybrid Holt’s Model with Wavelet-based ANN, tested against traditional ARIMA models and a Vanilla LSTM, integrated with a simple adjustment algorithm for daily or weekly forecasting.
• Results: The proposed model shows improved forecast accuracy over traditional models, providing more precise predictions for both the entire country and hotspot states.

Technical Details

Technological frameworks used: Hybrid Holt’s Model, Wavelet-based ANN, Vanilla LSTM, ARIMA, SIR model

Models used: Hybrid Holt’s Model with Wavelet-based ANN, Vanilla LSTM, ARIMA with and without wavelet-based functions

Data used: Daily confirmed COVID-19 cases from the entire country and 6 hotspot states.

Potential Impact

Healthcare sector, government agencies, epidemiological research firms, and public health analytics companies

Want to implement this idea in a business?

We have generated a startup concept here: PandemiPredict.